This visualization algorithm is very fast and efficient
and needs very little space, but it does need a little bit more
information from the scientific
computation part. Fortunately, the additional information already exists.
In this project, the author only finished the work on curvilinear (structured) grids with straight line boundaries. It needs to be expanded to more general cases, such as structures with curved boundaries, unstructured and hybrid grid cases. This should not be too difficult, because in FEM, the meshing theory for the other kinds of grids is almost the same as that of curvilinear grids. The shape functions already exist for all those cases. Deriving their derivatives and constructing the Newton-Raphson iteration formulas is easy. The only difficulty is that the algorithm needs to know how the regular space coordinates are formed in the unstructured case.